• by Erik Sass , Staff Writer @eriksass1, August 5, 2010

Everyone's talking about Google's rumored project to build a new social network, tentatively called "GoogleMe," speculating about Google's timing, approach, and chances of success -- the "when" and "how" of it. There's also a good amount of talk about its potential partners, including various social game makers, which covers the "who." But there's one question that's receiving rather less attention: why?

I for one am stumped: I can't see what Google expects to gain by launching and operating a new social network, even if it succeeds in toppling Facebook, or at least taking some of its market share (a big if, given Facebook's scale and momentum). The only hint I can find is a rather cryptic comment from CEO Eric Schmidt, who predicted in an interview earlier this week that "search is going to get better with more social information."

The meaning of "better" is vague, and probably boils down to "more accurate," "more popular," or "more profitable" (or all the above). But as far as I can tell, current trends and dynamics don't really support any of these predictions.

It would be great if combining social media and search made search "more accurate," in the sense of producing more relevant search results. There is indeed all kinds of searchable content on sites like Facebook, including profiles, messages, applications, music, video, status updates, photos, links, and so on. But there is no reason to think lots of people are trying to find this content via search. According to statistics from Alexa.com ( http://www.alexa.com/siteinfo/facebook.com ), less than 7% of Facebook's traffic (in terms of site visits) comes from external search engines, compared to around 50% for Wikipedia, 20% for Blogger.com, and 15% for YouTube.

So how about internal search on Facebook (including outbound search, bringing users from Facebook to the larger Internet)? If advertisers can use cookies to link search activity to social network profiles, they could combine search query data with information about the searcher's demographic characteristics, habits, and brand preferences, among other things, for more effective targeting of both search and display ads. But this assume trends in user behavior which have yet to materialize. By most accounts search is still a marginal activity on Facebook. In March U.S. users generated 647 million queries with Facebook search, or about five search queries per user in that period. That's a tiny fraction of the 570 billion page views generated by Facebook users that month, and it also pales in comparison to Google: in May, roughly 180 million unique U.S. visitors to Google's various sites generated 10.2 billion searches, according to ComScore, for an average 56.7 searches per visitor, equal to 63.7% of the search market.

In short, social network activity (at least on Facebook) appears to be dominated overwhelmingly by browsing -- a mostly undirected activity driven by idle curiosity, pretty much the opposite of search. Unless this changes for some reason, it's hard to see how social information can help make search "better," meaning either more accurate or more popular, since relatively few people are searching for content on social networks in the first place, internally or externally. And I haven't even mentioned Facebook's notoriously low click-through rates, or the fact that it still lacks an effective display advertising model.

All in all, it's kind of odd to watch Google, on course to take in over $26 billion this year, mightily exerting itself to compete with Facebook, which might see revenues of $1.4 billion. It's especially curious considering Facebook's growth to date doesn't seem to have diminished the usefulness of online search or the profitability of Google's search marketing business. But I'd be interested to hear readers' thoughts about what Google can achieve with social media -- especially profitable synergies or alignments I have overlooked.